• adult education
• basic skills assessment
• data analytics
• dati educativi
• educational data
• educazione degli adulti
• tecniche di analisi dei dati
• valutazione delle competenze di base

The EC publication “Horizon Report Europe: 2014 Schools Edition”, analyzed the emerging technologies and their impact on education. The report identified a number of trends and challenges to be faced by educational institutions in the near future. The educational data, made available thanks to the introduction of web-based learning sys-tems and the large scale adoption of ICT in education, allows the analysis of the named trends and challenges through different types of analytics. These include Educational Data Mining, Learning Analytics and Academic Analytics.

At the same time, the contemporary information society is mostly grounded on text-based technologies where reading, writing and numeracy skills play a fundamen-tal role. International, European and national organizations, recognizing the strategic importance of human capital for economic growth, have placed a lot of efforts in defin-ing core skills. OECD developed a strategy called “Better skills. Better jobs. Better lives” to help countries analyze their national skills systems. The most relevant program launched by OECD has been the Survey PIAAC which started in 2012 and was later on supported by a tool called EsOnline. Furthermore, in Italy, the national school reform identified the certification of basic skills as one of the priorities to be implemented.


This research intends to investigate the possible use of EsOnline precisely to re-spond to this need. Two case studies were developed, the first intended to test EsOnline in a real adult education setting, in Grosseto, Italy. A second case study was based on the data available through the 2012 PIACC Survey covering more than 30 countries including Italy. Data mining techniques were applied to the Italian dataset and results were used to build a classifier.

The analyses carried out with reference to the first case study allowed to answer several research questions: a first question was related to benchmarking local samples with national and international ones. We investigated whether local samples scored dif-ferently than the national and OECD ones. The question was answered positively since the local performance was found to be higher. A second research question was related to the change of core skills of adult students after attending a full school year. However, in this case, results were not conclusive. Research question 3 analyzed differences in the mean scores when considering the learning period at school. The analysis revealed that, on average, the students belonging to the second learning period scored signifi-cantly higher than those of the first one. A further research question was formulated on
whether EsOnline can be effectively contribute to the Individual Training Pact, and this was answered positively. Finally we investigated conditions for further adoption of Es-Online in the adult school system. A number of drivers and constraints were identified, which could support the adoption process in the future.

Regarding the second case study an analysis was performed on the datasets of the large-scale survey PIAAC, resulting in homogeneous sub-groups of test takers with common features. Visual analysis supported spatially the “digital division” of the sam-ples. Finally, the above data-driven process was applied to 28 OECD national datasets. The Italian dataset made it possible to build a classifier and each student in the Grosseto case was assigned to a cluster. Most of the test takers were assigned to the better per-forming clusters (where skilled workers prevail) and their average level of proficiency was consistent with the level computed for the clusters of the national dataset. The clas-sification procedure, without having the ambition of replacing EsOnline, could offer an effective and immediate support in the definition of the Individual Training Pact, by indicating a cluster membership and, associated to it, a level of basic skills proficiency.

A limitation of this study is that PIAAC and EsOnline datasets are not fully com-parable. This limits the possibility of predicting the level of performance with regards to basic skills. The use of the full set of PIAAC variables offers, in principle, important opportunities to be investigated by the communities of Educational Data Mining, Learning Analytics and Academic Analytics.

Investigating further the relation between skills performance, as defined and mea-sured by PIAAC and skill competences as established by the Italian Ministry of Education, would help evaluating competences at the different levels of the school system.